Decoding Machine Learning: Unveiling the Learning Process

Jagroop Singh - Oct 29 '23 - - Dev Community

Do you ever notice that when you go through shorts on YouTube, reels on Instagram, or videos on Tiktok, the type of content you liked or watched on these platforms is of similar kind that you like or you are interested on? 🧐📺

As a BTS Army, this happens to me all the time. 'I get BTS related videos all the time and I loved to watch it.' 💜🎶

This is not coming randomly but this is due to some trained algorithms that can automatically learn from our data and offer the best results possible. 🤖📊

This is only one example of Machine Learning; there are many others, such as self-driving cars 🚗🤖 and weather forecasting for the following days and months. ☀️🌧️🌦️🌤️

Thus,

Machine Learning has given the computer system abilities to automatically learns from data without being explicitly programmed.

Furthermore, it is not a one-step procedure in which your data is given to an algorithm and the output is provided; rather, it comprises a number of phases. 🔄🔍

So let's go over this one by one. 1️⃣👉

Process of Machine Learning


Step 1: Gathering Data

The most critical step is data collection. 📊📈 We can't go any further without it. For example, on Instagram, they will gather our data as well as the data of others. 📸🔍 Because we use these platforms, companies have no shortage of data, yet they still wish they had more. 📥💡 You know how we all want for more money, friends, and materialistic things? 💰👫🏢

Step 2 : Data Exploration

Engineers will investigate the data in this step. 🔍👩‍💼👨‍💼 They have a lot of data, and sometimes they have to be selective about the data in order to construct machine learning models. 🧐📊 For example, while watching reels, they will focus more on our age group, interests, liked, and remarked reels and focus less on data such as our qualifications, degree, and so on. 📆🎯📚🔘

Step 3 : Data Wrangling

There is really little likelihood that the data they collect is completely free of errors. 🚫📉 There are numerous outliers in it, including irrelevant, noisy, and missing data, all of which will reduce the precision of our machine learning model. 📊📉🤨 Therefore, the purpose of this stage is to clean the data so that the machine learning model can produce the most accurate results. 🧹🧽📈💡

Step 4 : Data Analysis

In this step, we will examine the algorithms we will employ based on the data. 🧐📊 Here, we determine the type of problem and prepare a strategy to acquire the best results from data. 🤖📈📚💡

Step 5 : Train Model

The model is now ready to be trained. 🚂📊 In this step, we train our model to increase its performance for better problem results. 📈📚

We train the model with datasets using multiple machine learning algorithms. 🤖📊 A model must be trained in order for it to learn the numerous patterns, rules, and features. 📝🧠🔍

Step 6 : Test Model

We test our machine learning model once it has been trained on a specific dataset. 🧪🤖 In this stage, we verify our model's accuracy by feeding it a test dataset. 📝📊

The percentage correctness of the model is determined by testing it against the project or problem requirements. ✔️📈

Step 7 : Deployment
The last step of the machine learning life cycle is deployment, where we deploy the model in a real-world system. 🚀🌐🤖


Wow, so many steps. Yes, but this is only a high-level overview of how it's being implemented. Each process has more subtasks if we delve further into it. 🕵️‍♂️📋

But it's worth it; I appreciate Tiktok, Youtube, and Instagram for suggesting BTS, Blackpink, NewJeans, BlackSwan, and TXT-related videos to me. 🎵💃🎤 Seeing their songs, video collaborations, and memes was a lot of fun. 😄📺📷🎶

. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
Terabox Video Player